A quick review The clustering problem: partition genes into - PowerPoint PPT Presentation

A quick review  The clustering problem:  partition genes into distinct sets with high homogeneity and high separation  Hierarchical clustering algorithm: 1. Assign each object to a separate cluster. 2. Regroup the pair of clusters with shortest distance. 3. Repeat 2 until there is a single cluster.  Many possible distance metrics  K-mean clustering algorithm: 1. Arbitrarily select k initial centers 2. Assign each element to the closest center • Voronoi diagram 3. Re-calculate centers (i.e., means) 4. Repeat 2 and 3 until termination condition reached

Biological Networks Analysis Introduction Genome 373 Genomic Informatics Elhanan Borenstein

Why we need networks (and systems biology)? VS.

Biological networks What is a network? What networks are used in biology? Why do we need networks (and network theory)? How do we find the shortest path between two nodes?

What is a network?  A collection of nodes and links ( edges )  A map of interactions or relationships

A long history of network/graph theory!! Network theory Graph theory Social sciences Computer science (and Biological sciences) Mostly 20 th century Since 18 th century!!! Modeling real-life Modeling abstract systems systems Measuring Solving “graph - structure & topology related” questions

The Seven Bridges of Königsberg  Published by Leonhard Euler , 1736  Considered the first paper in graph theory Leonhard Euler 1707 – 1783

The Seven Bridges of Königsberg  Published by Leonhard Euler , 1736  Considered the first paper in graph theory Leonhard Euler 1707 – 1783

Edge properties and special topologies  Edges:  Directed/undirected  Weighted/non-weighted  Simple-edges/Hyperedges  Special topologies:  Trees  Directed Acyclic Graphs (DAG)  Bipartite networks

Transcriptional regulatory networks  Reflect the cell’s genetic regulatory circuitry  Nodes : transcription factors and genes;  Edges: from TF to the genes it regulates  Directed; weighted?; “almost” bipartite  Derived through:  Chromatin IP  Microarrays  Computationally

Metabolic networks  Reflect the set of biochemical reactions in a cell  Nodes: metabolites  Edges: biochemical reactions  Directed; weighted?; hyperedges?  Derived through:  Knowledge of biochemistry  Metabolic flux measurements  Homology S . Cerevisiae 1062 metabolites 1149 reactions

Protein-protein interaction (PPI) networks  Reflect the cell’s molecular interactions and signaling pathways (interactome)  Nodes: proteins  Edges: interactions(?)  Undirected  High-throughput experiments:  Protein Complex-IP (Co-IP)  Yeast two-hybrid  Computationally S . Cerevisiae 4389 proteins 14319 interactions

Other networks in biology/medicine

Non-biological networks  Computer related networks:  WWW; Internet backbone  Communications and IP  Social networks:  Friendship (facebook; clubs)  Citations / information flow  Co-authorships (papers)  Co-occurrence (movies; Jazz)  Transportation:  Highway systems; Airline routes  Electronic/Logic circuits  Many many more …

The Bacon Number Game Tropic Thunder (2008) Tropic Iron Man Flatliners Thunder Proof Tom Cruise Robert Downey Jr. Gwyneth Paltrow Hope Davis Kevin Bacon Tropic Iron Man Thunder Frost/Nixon Tom Cruise Robert Downey Jr. Frank Langella Kevin Bacon

The Paul Erdos Number Game

The shortest path problem  Find the minimal number of “links” connecting node A to node B in an undirected network  How many friends between you and someone on FB (6 degrees of separation, Erdös number, Kevin Bacon number)  How far apart are two genes in an interaction network  What is the shortest (and likely) infection path  Find the shortest (cheapest) path between two nodes in a weighted directed graph  GPS; Google map

Dijkstra’s Algorithm "Computer Science is no more about computers than astronomy is about telescopes." Edsger Wybe Dijkstra 1930 – 2002

Dijkstra’s algorithm  Solves the single-source shortest path problem:  Find the shortest path from a single source to ALL nodes in the network  Works on both directed and undirected networks  Works on both weighted and non-weighted networks  Approach:  Iterative : maintain shortest path to each intermediate node  Greedy algorithm  … but still guaranteed to provide optimal solution !!

Dijkstra’s algorithm 1. Initialize : i. Assign a distance value, D, to each node. Set D to zero for start node and to infinity for all others. ii. Mark all nodes as unvisited. iii. Set start node as current node. 2. For each of the current node’s unvisited neighbors: Calculate tentative distance, D t , through current node. i. If D t smaller than D (previously recorded distance): D  D t ii. iii. Mark current node as visited (note: shortest dist. found). 3. Set the unvisited node with the smallest distance as the next "current node" and continue from step 2. 4. Once all nodes are marked as visited, finish.

Dijkstra’s algorithm  A simple synthetic network 2 B D 5 9 A F 1 4 3 7 9 3 C E 12 2 1.Initialize: i. Assign a distance value, D, to each node. Set D to zero for start node and to infinity for all others. ii. Mark all nodes as unvisited. iii. Set start node as current node. 2. For each of the current node’s unvisited neighbors: Calculate tentative distance, D t , through current node. i. ii. If D t smaller than D (previously recorded distance): D  D t iii. Mark current node as visited (note: shortest dist. found). 3.Set the unvisited node with the smallest distance as the next "current node" and continue from step 2. 4.Once all nodes are marked as visited, finish.

Dijkstra’s algorithm  Initialization  Mark A (start) as current node 2 D: ∞ D: ∞ A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ A F 1 4 3 7 9 D: 0 D: ∞ 3 C E 12 D: ∞ D: ∞ 2

Dijkstra’s algorithm  Check unvisited neighbors of A 2 0+9 vs. ∞ D: ∞ D: ∞ A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ A F 1 4 3 7 9 D: 0 D: ∞ 3 C E 12 D: ∞ D: ∞ 2 0+3 vs. ∞

Dijkstra’s algorithm  Update D  Record path 2 D: ∞ D: ∞ ,9 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 D: 0 D: ∞ 3 C E 12 D: ∞ ,3 D: ∞ 2

Dijkstra’s algorithm  Mark A as visited … 2 D: ∞ D: ∞ ,9 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 D: 0 D: ∞ 3 C E 12 D: ∞ ,3 D: ∞ 2

Dijkstra’s algorithm  Mark C as current (unvisited node with smallest D) 2 D: ∞ D: ∞ ,9 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 D: 0 D: ∞ 3 C E 12 D: ∞ ,3 D: ∞ 2

Dijkstra’s algorithm  Check unvisited neighbors of C 3+3 vs. ∞ 2 3+4 vs. 9 D: ∞ D: ∞ ,9 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 D: 0 D: ∞ 3 C E 12 D: ∞ ,3 D: ∞ 3+2 vs. ∞ 2

Dijkstra’s algorithm  Update distance  Record path 2 D: ∞ ,6 D: ∞ ,9,7 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 7 3 6 5 ∞ D: 0 D: ∞ 3 C E 12 D: ∞ ,3 D: ∞ ,5 2

Dijkstra’s algorithm  Mark C as visited  Note: Distance to C is final!! 2 D: ∞ ,6 D: ∞ ,9,7 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 7 3 6 5 ∞ D: 0 D: ∞ 3 C E 12 D: ∞ ,3 D: ∞ ,5 2

Dijkstra’s algorithm  Mark E as current node  Check unvisited neighbors of E 2 D: ∞ ,6 D: ∞ ,9,7 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 7 3 6 5 ∞ D: 0 D: ∞ 3 C E 12 D: ∞ ,3 D: ∞ ,5 2

Dijkstra’s algorithm  Update D  Record path 2 D: ∞ ,6 D: ∞ ,9,7 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 7 3 6 5 ∞ D: 0 D: 0 D: ∞ ,17 7 6 5 17 3 C E 12 D: ∞ ,3 D: ∞ ,5 2

Dijkstra’s algorithm  Mark E as visited 2 D: ∞ ,6 D: ∞ ,9,7 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 7 3 6 5 ∞ D: 0 D: ∞ ,17 7 6 5 17 3 C E 12 D: ∞ ,3 D: ∞ ,5 2

Dijkstra’s algorithm  Mark D as current node  Check unvisited neighbors of D 2 D: ∞ ,6 D: ∞ ,9,7 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 7 3 6 5 ∞ D: 0 D: ∞ ,17 7 6 5 17 3 C E 12 D: ∞ ,3 D: ∞ ,5 2

Dijkstra’s algorithm  Update D  Record path (note: path has changed) 2 D: ∞ ,6 D: ∞ ,9,7 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 7 3 6 5 ∞ D: 0 D: ∞ ,17,11 7 6 5 17 3 C E 12 7 6 11 D: ∞ ,3 D: ∞ ,5 2

Dijkstra’s algorithm  Mark D as visited 2 D: ∞ ,6 D: ∞ ,9,7 A B C D E F B D 5 9 0 ∞ ∞ ∞ ∞ ∞ 0 9 3 ∞ ∞ ∞ A F 1 4 3 7 9 7 3 6 5 ∞ D: 0 D: ∞ ,17,11 7 6 5 17 3 C E 12 7 6 11 D: ∞ ,3 D: ∞ ,5 2